Maximum drawdown — what does it mean for your account?

In a hypothetical example, an account rises from EUR 10,000 to EUR 11,500, then falls to EUR 8,700. The decline from the earlier peak is 2,800 ÷ 11,500 ≈ 24.35%. This records a past decline; it does not establish a strategy’s quality, a future loss ceiling or a psychological threshold. The calculation depends on which capital series you measure.

What exactly does drawdown mean?

Drawdown measures a decline from a previous peak in a specified series: balance, equity, or an investment-return series adjusted for external cash flows. For a positive peak H and a later value K, the percentage is (H − K) ÷ H × 100%. State the period, observation frequency and series before comparing results.

drawdown = (peak − current equity) ÷ peak × 100%

Take a concrete case. The account starts at 10,000 USD. After two weeks equity reaches 11,500 USD — a new peak. A losing streak then drags it down to 9,800 USD, so the drawdown is (11,500 − 9,800) ÷ 11,500, or 14.8 percent. When the trough deepens to 8,700 USD, the drawdown grows to 24.3 percent.

With no external cash flows, an account that grows from USD 10,000 to USD 20,000 and then falls to USD 15,000 has a 25% drawdown, although it remains 50% above its start. Deposits and withdrawals need separate treatment. In a hypothetical unit model, USD 10,000 buys 100 units at USD 100. A fall to USD 8,000 lowers the unit value to USD 80. A deposit of USD 8,000 then buys another 100 units: the account reaches USD 16,000 but the unit value remains USD 80. A subsequent 10% fall leaves USD 14,400 and 200 units at USD 72. The unit series 100 → 80 → 80 → 72 has MDD 28%; the raw account series 10,000 → 8,000 → 16,000 → 14,400 shows only 20%. The deposit did not recover the investment loss. This simplified model requires valuations at the flow time and makes no claim of GIPS compliance.

What is maximum drawdown and why bother measuring it?

Maximum drawdown (MDD for short) is the deepest drawdown recorded in the period under review. If over twelve months the account suffered four slides from successive peaks — 8, 14, 23 and 11 percent — the maximum drawdown is 23 percent. Not the average of those four numbers, but the single worst episode. People confuse the two surprisingly often; average drawdown is a separate measure and usually a far gentler one.

MDD records the deepest observed decline under the stated measurement rules. Two hypothetical strategies may both return 20% over a year while recording MDD of 5% and 35%; their past paths differ. MDD alone does not provide a ruin probability, identify a profitable method, or determine what a person can tolerate. Read it alongside exposure, costs, drawdown duration and the assumptions behind the data.

Why is recovering a loss asymmetric?

The most merciless property of drawdown: getting back to where you started requires a larger percentage gain than the fall itself. That is arithmetic, not a matter of opinion — after a loss you are counting percentages off a smaller base.

The formula is required gain = d ÷ (1 − d), where d is drawdown as a fraction and 0 ≤ d < 1. Multiply the result by 100 for a percentage. A 50% drawdown needs 100% growth; an 80% drawdown needs 400%. After a 100% loss, the remaining zero cannot reach the old peak through any finite percentage return; new capital is a separate cash flow.

A hypothetical constant net return of 15% per year would double remaining capital in ln(2) ÷ ln(1.15) ≈ 4.96 years under a smooth compounding model. With only annual observations, the first year-end above double is year five. This explains the arithmetic after a 50% loss; it is not an attainable-return promise or an estimate of actual recovery time.

What should accompany the drawdown number?

A drawdown figure needs a measurement record. These checks concern what the number represents, not a diagnosis of trader reactions or a universal acceptable percentage.

No universal drawdown percentage guarantees that a strategy is suitable or survivable. Set an affordable review or stopping threshold beforehand, inspect the causes and current validity of the method, and recognize that execution gaps can bypass an intended threshold. A previously observed drawdown does not establish the future maximum.

How to review the measurement and the plan

1. Choose and record a sampling frequency. Keep the date, measured capital and highest prior capital under the same definition. End-of-week observations can miss an intraday trough; changing the frequency changes what the statistic can capture.
2. Review exposure and the stated loss budget. The 1 percent of capital per trade, drop to 0.5 percent example is a chosen sizing model, not a universal response to 10% drawdown. Partially offsetting the exposure can change sensitivities, costs and margin; it does not guarantee a smaller future drawdown.
3. Separate observed causes from assumptions. In the journal, distinguish rule-compliant trades, execution differences and rule deviations. That classification helps frame questions; it does not prove that a method has an edge or that emotions were the sole cause.
4. Use simulation to inspect a specified change. State the rules, data, costs and execution assumptions being tested. A set of demo trades can test whether the procedure was followed; it cannot certify future live performance or a maximum drawdown.
5. Document any review or stopping rule beforehand. A chosen threshold is a plan input, not a guaranteed cash-loss limit. Black-swan events and price gaps can lead to a worse fill than the intended stop price. Record how actual execution and open exposure affect the result.